Re: EXP

"Nicolas Bronke" <bronke_at_gmx.net> wrote in message news:<aibjr5$1349fm$1_at_ID-121870.news.dfncis.de>...
> > "depending order" ? Check the manual for consistent = y and see if
> > that is what you mean.
> >
> Yes and no. This was what I remembered but is actually not what I am looking
> for.
>
> As a better explanation:
> a)
> view x depends on view y
> now you exporting the schema and trying to import on another database. There
> you will get a failure. The result is an invalid view.
> This is a big problem, because I can recompile the view.
>
> b)
> As I found in some databases, there are views like
>
> select NewColumn,TableName.* from TableName ....
>
> With this I found out the problem, that the view is not imported in some
> cases. If I start the import twice, then mostly all objects are imported.
>
> Especially b) I wanr to avoid. I cannot change the view-source, because it
> is not from mine.
>
> I hope now I explained my problem a little bit better.
>
> Regards
> Nicolas

I think I understand what you want better, but if you are importing user A and A has a view that references a user B object and this user B object is not in the new instance there is no command to cause user B's object and the grants to A on it to import.

In the case of a full import Oracle brings the view late in the import so the tables and grants should be there plus it uses the force option to cause the source code to be imported even if the view compile fails so that when the missing object or grant is provided the code can just be recompiled.

I would suspect in a user import that Oracle still brings the source for views even if referenced objects or grants on then are missing, but I think it is up to you to find and import all the necessary non-owned objects.

HTH -- Mark D Powell -- Received on Fri Aug 02 2002 - 09:28:56 CDT